Search

The main objective of this dissertation was to utilize a molecular approach combining DFT calculations and numerous experimental techniques to develop carbon tolerant electrocatalysts for solid oxide fuel cell (SOFC) anodes. SOFCs are solid-state electrochemical devices that can convert the chemical energy of hydrogen, CO, and hydrocarbons into electrical energy. One of the main issues associated with the direct operation of SOFCs using hydrocarbons is the deactivation of the conventional anode electrocatalysts, such as Ni on yttria-stabilized zirconia (YSZ) due to the formation of carbon deposits. To tackle the problem of carbon-induced deactivation of Ni, we have utilized DFT calculations to identify the chemical transformations that govern carbon poisoning of Ni. We found that the processes of C-C and C-O bond formation as well as carbon nucleation played an important role in the carbon-induced deactivation of the Ni catalysts. These insights led to the identification of the Ni surface alloys as promising carbon tolerant catalysts. Electrocatalysts containing Sn/Ni and Ni supported on YSZ were synthesized. The carbon tolerance of the Sn/Ni surface alloy and monometallic Ni was tested using reactor and electrochemical SOFC studies. We found that the Sn/Ni surface alloy exhibited a significantly improved carbon tolerance compared to monometallic Ni in hydrocarbons steam reforming reactions and as SOFC anode electrocatalyst. To advance our understanding of the chemistry that occurs on the Sn/Ni surface alloy, we have also performed kinetic studies in combination with DFT calculations to identify the critical elementary steps that govern the performance of Ni and Sn/Ni electrocatalysts. We have also utilized electron microscopy and spectroscopy to measure the electronic structure of the supported Ni and Ni alloys and have related that to their chemical and catalytic performance. These detailed atomistic studies led to the derivation of very general set of principles that allow us to identify novel carbon tolerant alloy catalysts for hydrocarbon reforming and electro-oxidation. The work described in this dissertation presents a rare example where DFT calculations combined with experimental techniques led to the bottom-up (based on molecular insights rather than on empirical trial and error testing) identification of improved electrocatalysts. more...

The ability to selectively functionalize complex molecules is an ongoing challenge to chemists; the use of transition-metal catalysts can allow for the direct transformation of C–H bonds into a variety of different functional groups. Two significant challenges to this approach are regioselectivity and diversity of functionalization. The approach discussed herein towards achieving regioselectivity has been to use substrates with pendant ligands, which coordinate to a palladium catalyst and direct electrophilic C–H activation to a specific C–H bond. In order to incorporate a variety of functional groups into the substrate, both oxidative functionalization reactions, to form C–OR and C–F bonds, and oxidative coupling reactions, for the formation of a new C–C bond from two C–H bonds have been developed. We have pioneered a method for the Pd-catalyzed acetoxylation of activated and unactivated, primary and secondary sp3 C–H bonds using pyridine and quinoline directing groups. Additionally, we have demonstrated one of the first metal-catalyzed C–F bond-forming reactions, using commercially available electrophilic fluorine sources as oxidants. This methodology was applied to the formation of a both sp2 and sp3 C–F bonds. The formation of C–C bonds is an ongoing challenge to organometallic chemists. We felt that in order to maximize atom economy and the overall efficiency, the coupling of two C–H bonds to form a C–C bond would be an ideal process. Using heterocycles as directing groups, select C–H bonds can be coupled under mild reaction conditions. With Oxone as the terminal oxidant, Pd(OAc)2 catalyzes oxidative homodimerization of these heterocyclic substrates to afford homocoupled products in high yield and very high regioselectivity. The selective formation of unsymmetrical aryl-aryl’ bonds would be a more general process. We have found that palladium, in the presence of benzoquinone and Ag2CO3, can catalyze the cross-coupling of C–H bonds. The first C–H activation is ligand directed followed by a second, undirected C–H activation. The coupled products are formed in high regioselectivity, where the undirected C–H bond activation typically occurs at the least sterically hindered C–H bond. more...

The application of alkyne metathesis in the development of polymers, pharmaceuticals, natural products, and many other products has surged. In response, an expansion of current catalysts and methods to synthesize the catalysts is desired. This work focuses on the systematic design and development of tungsten and molybdenum-based alkylidyne and nitride complexes for triple-bond metathesis. Nitrile-alkyne cross metathesis (NACM), a new type of metathesis involving the exchange of alkyne and nitrile moieties was developed. NACM activity was found with three tungsten nitride complexes, [N≡W(OCMe2CF3)3]3 (1), N≡W(OCMe¬(CF3)2)3(DME) (2), and [Li(DME)2][N≡W(OCMe(CF3)2)4] (3). Optimization of the NACM conditions including reaction medium, temperature, concentration, and catalyst loading were completed. NACM resulted in a mixture of symmetrical and unsymmetrical alkyne products with 2 favoring the formation of symmetrical alkynes relative to 1. These product ratio differences were accounted for through variances in catalyst resting state during NACM. The selective formation of symmetrical or unsymmetrical alkynes in high yield can be achieved by altering the reaction conditions. Although 2 exhibits greater NACM activity than 1, broader substrate compatibility is found with 2. Catalyst deactivation was substrate-dependent; conversion to alkylidyne complexes prior to deactivation occurring with some substrates. The utility of NACM was illustrated through the synthesis of large arylene-ethynylene macrocycles. Methods to readily synthesize alkylidyne species from N≡Mo(OR)3 and Mo2(OR)6 were developed. The irreversible formation of benzylidyne complexes from N≡Mo(OCMe(CF3)2)3 and N≡Mo(OC(CF3)3)3(MeCN) was achieved. The first evidence for the formation of RC≡Mo(OR)3 complexes (OR=OCMe3, OCMe2CF3, and OCMe(CF3)2) through the interaction of Mo2(OR)6 complexes with internal alkynes was discovered. The first example of the reversible formation of alkylidyne and Mo2(OR)6 species (OR=OCMe3, OCMe2CF3) was also achieved. Extension of the formation of molybdenum alkylidyne complexes from nitride complexes that contain OCMe¬3 and OCMe2CF3 ligands was achieved in situ in the presence of Lewis acids. Attempted isolation of alkylidyne complexes formed in situ was unsuccessful due to decomposition of the formed alkylidyne species in the presence of the Lewis acid. Several simple Lewis acids were found to increase the rate of alkyne metathesis with N≡Mo(OR)3 and Mo2(OR)6 complexes (OR= OCMe3, OCMe2CF3, and OCMe(CF3)2). more...

We performed Supercritical Water Gasification (SCWG) for cellulose and lignin for the first time in the absence of catalytic effects from metallic walls, by using quartz reactors. We quantified the catalytic effect of metals by adding metal wires to the reactors. We also performed the first systematic study of the effect of variables on gas yields. We varied time (2.5 to 75 minutes), temperature (365-725C), water density (0.00-0.18 g/cm3), and biomass loading (1.0 wt %-33.3 wt %). In the absence of metals, high temperatures and water densities provide the highest gas yields. Up to 3.3 mmol/g of H2 were obtained from cellulose (at 0.18 g/cm3) and up to 7.5 mmol/g (at 725C) from lignin. Up to 2.6 mmol/g of CH4 were obtained from cellulose (at 600C), and up to 9.0 mmol/g from lignin (at 725C). The highest energetic content (LHV gas/LHV biomass) was 20.0 % from cellulose (at 600C), and 37.4 % from lignin (at 725C). The presence of metals increases gas yields to a significant extent if the catalyst surface area/biomass weight ratio is at least 15.4 mm2/mg (5.0 wt % biomass loading). Nickel and copper provide high yields at 5.0 wt % loading, and nickel provides the highest yields at 1.0 wt % loading. Nickel at 240 mm2/mg provides 23.5 mmol/g of H2 from cellulose and 21.1 mmol/g of H2 from lignin, which is close to equilibrium yields. CH4 yields are not strongly influenced by the presence of metals. With nickel, it is possible to generate a gas with almost 50 % energetic content from cellulose, and 45 % from lignin. The non-catalytic results were used to fit the first kinetic model describing gas formation in SCWG. The proposed 11 reactions and the concept of a generic intermediate led to a model that can successfully fit the base case experimental data for cellulose and lignin. We verified that the model can predict yields at different biomass loadings and water densities. Its equilibrium predictions agree with thermodynamic calculations, and the rate constants obtained for the water-gas shift are in the range reported by previous authors. more...

This thesis focuses on III-Nitrides for electronic and gas sensing devices. Systematic material growth optimization was performed by using various characterization techniques in order to improve device performance. For high frequency device operation, resistive bulk GaN is preferable to prevent parasitic leakage through the layer. Growth conditions for bulk GaN were optimized for this purpose using an in-house Metal Organic Vapor Phase Epitaxy system. The pinch-off current of AlGaN/GaN Heterostructure Field Effect Transistors based on these layers was reduced by a factor of twenty in comparison with devices fabricated with non-optimized bulk GaN. The best results obtained from this heterostructure were a room temperature Hall mobility and sheet charge density of about 1270cm2/Vs and 7.0•10^12cm−2, respectively. AlN/GaN HFETs were investigated because of the possibility of reducing gate leakage current, operating at higher temperature and higher power in comparison with AlGaN/GaN HFETs. In-situ SiNx was employed for surface passivation. SiNx improved the ohmic quality of this device and the RF characteristics were also improved; fT and fmax were enhanced by a factor of two and three, respectively in contrast to devices without the passivation. To address the issue pertaining to the polarization fields of c-plane GaN, non-polar GaN growth was investigated using r-plane sapphire substrates. The RMS surface roughness and full width at half maximum were improved to 2nm and 1000", respectively, using high temperature AlN buffer layers. This shows a quality improvement by a factor of two to three compared to layers grown using low temperature GaN NLs in this work. Diodes were fabricated for gas sensing using c-plane GaN and AlGaN/GaN heterostructures. Optimization of size and thickness of the Pt Schottky contact improved CO sensitivity by a factor of six compared to non-optimized sensors. Fabry-Pérot filters with GaN/air gap based distributed Bragg reflectors were designed at 450nm as a detector in optical gas sensing systems. Simulations of their optical and mechanical properties showed the feasibility of this device design. The growth and etching study of AlN as a sacrificial layer manifested that reasonable etching rate can be obtained when the layer was grown at 800C. more...

Pure monosilicic acid solutions mimic the precipitation of silica dissolved from analcime in HCl. Thus, mineral dissolution and silica precipitation are decoupled and pure monosilicic acid solutions were used to study the precipitation of silicon dissolved from minerals. Monosilicic acid solutions were used to study silicon precipitation using UV-Vis and DLS. UV-Vis results showed that under very acidic conditions monosilicic acid disappears rapidly from solution and follows second-order disappearance kinetics. Then, from DLS, it is shown that within the first few minutes, monosilicic acid polymerized to form primary particles ~5nm in diameter, which then flocculate, with mean floc diameter increasing exponentially with time. Both the monomer disappearance and flocculation rate increase with increasing HCl concentration. The flocculation was computer-simulated using a Smoluchowski equation modified for a geometric population balance with a reaction-limited aggregation (RLA) kernel. DLS-measured mean silica particle size agreed well with the simulated mean particle sizes. Salt experiments indicate that 1M salt added to HCl can increase monosilicic acid disappearance rate and silica particle growth rate in the order: AlCl3 > CaCl2 > MgCl2 > NaCl > CsCl > no salt. UV-Vis results showed that monosilicic acid disappears rapidly and follows a third-order kinetic rate law in contrast to second-order when no salt is present. DLS results showed that silica particle size also increases exponentially with time when salt is present. It was also found that specific reaction rate constants: monosilicic acid disappearance rate constant, particle growth rate constant and computer-simulation aggregation rate constant, increase exponentially for all salts as a function of solution ionic strength. Finally, initial dissolution rate of analcime in the strong acids HCl, HBr, and HNO3 are very similar and follow a Michaelis-Menten mechanism. At proton concentrations less than 0.055M, initial dissolution rate in citric acid was faster than initial dissolution rate in the strong acids. The observed plateau in AAS/ICP-MS silicon concentration during analcime dissolution exists because of a recondensation reaction between silanol groups within dissolving analcime. The recondensation rate varies in strong acids corresponding to the order: H2SO4 > HI > HBr > HCl > HNO3 more...

Olefin metathesis (OM) has become a widely used tool in organic and material syntheses. As catalyst development has advanced, functional group tolerance has increased. Unfortunately, vinyl halides were incompatible with OM catalysts and attempts at cross-metathesis (CM) with vinyl halides failed. Given the usefulness of alkenyl halides in metal-catalyzed cross-coupling reactions, improvement of CM systems employing vinyl halides would be beneficial. Research goals included determining why vinyl halides were not tolerated by ruthenium-based OM catalysts and developing systems in which vinyl halides participate in CM. Ruthenium monohalomethylidene complexes, [Ru(=CHX)(H2IMes)LCl2] (X = F, Cl; L = PCy3, 2 py), which are intermediates in CM with vinyl halides, were found to undergo decomposition through loss of HX, deactivating the catalyst. Decomposition of the monochloromethylidene complexes can be hindered by removing one of the neutral ligands (L) from the system, making CM with chlorinated olefins successful. The monofluoromethylidene complexes synthesized were less susceptible to decomposition. However, the monofluoromethylidene intermediates act as a thermodynamic well, shutting down CM. If metathesis product formation is energetically favored, then the thermodynamic stability of the monofluoromethylidene intermediate relative to Ru-alkylidene intermediates can be overcome by this added driving force. Therefore, vinyl fluoride is an effective substrate for ring-opening cross-metathesis with cyclooctene and enyne metathesis with a number of alkynes. We also pioneered a new subfield of OM referred to as Fischer carbene cross-metathesis (FCM), in which the thermodynamic stability of Fischer carbene complexes was circumvented by removing the need to form a ruthenium alkylidene complex during the catalytic cycle. FCM involves CM of electron-rich olefins such as ethyl vinyl ether with functionalized 1,2-disubstituted alkenes such as styryl acetate to form β-ethoxystyrene. FCM has allowed for productive CM with a number of electron-rich olefins including vinyl fluoride which have previously been detrimental to Ru-based catalysts. Finally, the decomposition process of the monohalomethylidene complexes led to the discovery of a facile synthesis of new Ru-benzylidyne species. The synthesis and reactivity of a number of previously unknown Ru-benzylidyne complexes was studied and has important implications for the development of Ru-based alkyne metathesis catalysts, which have yet to be realized. more...

The increased use of biofuels presents an opportunity to improve combustion performance while simultaneously reducing greenhouse gases and pollutant emissions. This work focused on improving the fundamental understanding of the auto-ignition chemistry of oxygenated reference fuel compounds. A systematic study of the effects of ester structure on ignition chemistry was performed using the University of Michigan Rapid Compression Facility. The ignition properties of the ester compounds were investigated over a broad range of pressures (P=5-20 atm) and temperatures (T=850-1150 K) which are directly relevant to advanced combustion engine strategies. Ignition delay times for five esters were determined using the RCF. The esters were selected to systematically consider the chemical structure of the compounds. Three esters were saturated: methyl butanoate, butyl methanoate, and ethyl propanoate; and two were unsaturated: methyl crotonate and methyl trans-3-hexenoate. The unsaturated esters were more reactive than their saturated counterparts, with the largest unsaturated ester, methyl trans-3-hexenoate having the highest reactivity. Two isomers of the saturated esters, butyl methanoate and ethyl propanoate, were more reactive than the isomer methyl butanoate. The results are explained if we assume that butyl methanoate and ethyl propanoate form intermediate ring structures which decompose more rapidly than esters such as methyl butanoate, which do not form ring structures. Modeling studies of the reaction chemistry were conducted for methyl butanoate and ethyl propanoate, for which detailed mechanisms were available in the literature. The new experimental data indicated that literature rate coefficients for some of the methyl butanoate/HO2 reactions were too fast. Modifying these within the theoretical uncertainties for the reaction rates, led to excellent agreement between the model predictions and the experimental data. Comparison of the modeling results with the intermediates measured during methyl butanoate ignition indicated that pathways leading to the formation of small hydrocarbons are relatively well represented in the reaction mechanism. The results of this work provide archival benchmark data for improved understanding of the dominant reaction pathways and species controlling the auto-ignition of oxygenated reference fuel compounds. These data also provide a path for continued development of chemical kinetic models to optimize practical combustion systems. more...

In this dissertation, liquid feed–flame spray pyrolysis (LF-FSP) was used for the synthesis and characterization of mixed-metal oxide nanopowders in the Ce-Zr-Al-O system for potential catalytic, photonic, electronic and structural applications. The LFFSP process aerosolizes metalloorganic precursors dissolved in an alcohol solvent with oxygen, combusts the aerosol, and then rapidly quenches the gaseous species to produce nanosize oxide “soot” with the same compositions as those in the starting precursor solutions. The wide variety of potential metal oxide precursors allows the synthesis of multiple chemical compositions, phase and unique particle morphologies in the resulting nanopowders. Surprisingly, LF-FSP with the correct choice of metalloorganic precursors provides access to core-shell nanopowders in a single step. The (ZrOx)x(Al2O3)1-x system was explored first for potential structural applications as zirconia toughened alumina (ZTA) composites. During the LF-FSP process, core-shell nanoparticles with more...

Realizing the need for effective kinetic models that could be used over wide operating regimes, oxidation and storage kinetics for a diesel oxidation catalyst (DOC) were developed in this work. As a first step towards kinetics development, a simple catalyst formulation including only Platinum was chosen. Kinetics were generated by assuming that propylene was representative all the hydrocarbons (HCs) in the exhaust. A systematic methodology was formulated which consisted of (1) careful choice of concentration/temperature domain (2) measurement of reactor conversions of aged catalyst samples at chosen test points using a high space velocity integral reactor (3) developing a simplified 1D reactor model (4) defining an objective function which is critically sensitive to the differences between model predictions and experiments at all conversions (5) generating proper initial guesses and finally (6) modifying Langmuir-Hinshelwood rate expressions to arrive at the final rate forms. This methodology can be used to generate steady state global kinetics in general. Comparison of model predictions with light-off curves generated using a 1.7L Isuzu diesel engine revealed that propylene is not representative of all the HCs in the diesel exhaust. As a next step towards oxidation kinetics development, a commercially available DOC catalyst was used with HCs in the diesel exhaust speciated as propylene, representing partially oxidized HCs, and diesel fuel, representing unburnt fuel component in diesel exhaust. The systematic methodology developed previously was successfully used to generate oxidation kinetics for all the species of interest. Light-off curves comparison revealed excellent agreement between model predictions and engine data. Finally, reaction kinetics were developed for capturing hydrocarbon adsorption/desorption processes on zeolite. For this study the fuel components in the exhaust were further speciated as n-dodecane and toluene. A minimum of four experiments were found to be sufficient to generate the necessary kinetic constants for each adsorbable HC species. Studies on simplified warm-up process using a 1D adiabatic reactor model that incorporated both the oxidation and storage kinetics indicated that the storage component reduces the overall cold start HC emissions by at least a factor of 2 if the warm-up rate achieves 45-65C/min, a range commonly observed during start-up. more...

In this work, we developed a novel power supply for the WIMS-ERC intraocular sensor (WIMS-IOS), an autonomous and implantable system. This device is representative of a broad class of microscale devices, whose full implementation in environmental and medical systems will require significantly smaller power supplies; presently, battery systems represent 85% mass and 50% volume of typical devices. Strategies using both commercial and specially developed devices, using a variety of electrochemistries have been used. The smallest of the batteries reported to date, are thin-film lithium (Li) cells, using a chemical vapor deposition (CVD) or a pulsed laser deposition (PLD) approach. Thin film batteries fabricated with these techniques have achieved electrodes thicknesses less than 5µm (unpackaged), capacities of ~100µAh/cm2 and footprints of 1cm2. However, clean-room fabrication and high power laser equipment needed for ceramic materials entail high cost (~$300/Wh), and the elevated processing temperatures (500-720ºC) and use of chemicals and etchants, make them incompatible with CMOS materials. Finally, the intrinsically high power (3.5-4.2V) of lithium chemistry complicates integration with low-voltage MEMS, since it necessitates voltage regulation. In our study we deposited thin film electrodes using physical vapor deposition (PVD), a low temperature (270-500ºC) purely physical process in a vacuum ~10-7Torr. Our underlying hypothesis was that this technique would reduce intrinsic losses because of the high resulting precision, while allowing integration with chips because of more benign processing conditions to MEMS. Our specific objectives were to: 1) analyze commercial systems for the WIMS-IOS; 2) create batteries from commercial active materials; and finally 3) create test and integrate novel batteries. Commercial Zn/Ag batteries were selected using a previously developed system analyzer (POWER algorithm). Active materials from the same commercial systems were packaged in ceramics and tested in three cells. Six novel cells (cathode/anode) were deposited on MEMS platforms (glass substrates). The last of these was integrated with the WIMS-IOS microchip and tested. Deposited batteries had suitable voltages (1.55) for MEMS, thicknesses of ~25µm (unpackaged) and footprints of ~2mm2, with capacities ~0.1mAh/cm2, and processing conditions compatible with MEMS materials. They also offer potentially lower cost than existing systems. more...

This dissertation examines the ability of nickel-based catalysts to reform liquid hydrocarbons into reformate suitable for solid oxide fuel cell auxiliary power units and automotive emission treatment strategies. n-Dodecane and tetralin served as model compounds representing hydrocarbons in jet fuel. During autothermal reforming, tetralin showed the highest conversion, greatest product yields, and higher reaction temperature with the lowest amount of carbon deposition and fewer cracking products than n-dodecane. Carbon deposition during autothermal reforming was strongly dependent on the amount of smaller hydrocarbons produced during the reaction. Monolith catalysts, suitable for mobile application, were prepared to elucidate the role of mixed oxide supports and nickel during reforming reactions; catalysts contained nickel metal supported on cerium-zirconium oxide deposited on cordierite monolith and nickel deposited directly on cordierite. The catalysts were studied under autothermal reforming, partial oxidation, and steam reforming conditions. Nickel acted as a site for both partial oxidation as well as steam reforming. Nickel loadings above 2 wt.% showed marginal gains in reformate yield. Cerium-zirconium oxide as a support for nickel increased the participation of water in the autothermal reforming reaction, increased nickel dispersion, and acted as a site for hydrocarbon oxidation. Homogeneous cracking reactions were a significant source of n-dodecane conversion, especially at elevated temperatures. Steam reforming of the liquid hydrocarbon cracking products was an important step in the production of reformate. Temperature programmed oxidation of post reforming catalysts showed the presence of ∼600°C carbon associated with nickel and ∼500°C carbon associated with cerium-zirconium oxide. The ∼600°C carbon was the most damaging to the monolith structure. The rate of carbon deposition increased with larger nickel particle sizes. Cerium-zirconium oxide decreased the rate of carbon deposition. Catalyst deactivation was related to the formation of filamentous carbon that separated the nickel from the cerium-zirconium oxide and disintegrated the monoliths. Overall, 2 wt.% nickel supported on cerium-zirconium oxide was the best catalyst for the autothermal reforming of liquid hydrocarbons, resulting in hydrogen yields of up to 120% with carbon deposition under 2 wt%. more...

Catalysts comprising nickel supported on a reducible ceria-zirconia mixed oxide were prepared with a range of nickel loadings. Supports were synthesized by coprecipitation, calcined in air, and impregnated with nickel. The effects of nickel loading on catalytic performance for the autothermal reforming of isooctane were investigated. It was found that there is an interaction between nickel and the reducible oxide support. The effects of this interaction are most strongly evident at low nickel loading, and include modification of the reduction temperatures of both the nickel and support components, and significant increase in resistance to carbon deposition under autothermal reforming reaction conditions. At higher nickel loadings the carbon resistance is lost, and reaction at low H2O/C and O/C ratios leads to heavy coke deposition. The exact nature of the nickel-support interaction in these catalysts is not clear from the current investigation. It is possible that the effects are due to limited incorporation of nickel into the support lattice, thereby introducing oxygen defects and new types of catalytic sites, or it may simply be an effect of interfacial contact between small nickel domains and the support, resulting in an enhanced rate of oxygen delivery to nickel during reaction. The lowest nickel loading studied, 1 wt%, yielded catalysts that had stable operation under demanding autothermal reforming conditions. At a H 2O/C and O/C of 0.5, high activity was maintained over the course of eight hours on stream. Given the propensity of nickel catalysts to form carbon during reforming of hydrocarbons, this result is novel and unexpected. It offers the opportunity to develop fuel processing technology around catalysts which do not use precious metals, and can operate with feed stoichiometries previously thought unusable. In an effort to develop an integrated fuel processor based on these catalysts, they were tested as washcoats in prototype microchannel reactors. Catalyst performance in the microchannel reactors was generally poor as compared to the same catalysts tested in a packed bed configuration. This seems to be attributable to the differences in heat transfer between the two reactor materials and geometries, leading to different temperature profiles in the reactors. This indicates that reactor design in microchannel systems will need to be adjusted for reactions with high exotherms or endotherms. more...

The purpose of this research effort was two fold. The first objective was to understand the process and material issues relating to producing mixed ionic and electronic conducting electrodes (MIECE) made from heat-sintered blends of molybdenum and titanium dioxide, and from heat-sintered blends of titanium nitride and titanium dioxide. The second objective was to investigate the fundamental physical characteristics, such as surface contaminants and morphology at the electrode/electrolyte interface, to determine how those characteristics influences the mechanisms and energies which govern the oxidation and reduction reactions at the electrodes. The intent was to learn what were the contributing components to electrical impedances and overpotentials, to measure them directly, or indirectly, and to control them for optimization. Sodium beta″-alumina was used as the solid electrolyte ionic conductor with sodium as the mobile species. The major processes and equipment used in this research program were presented, as were the experimental methods and techniques employed to evaluate the results. While the efforts to obtain a MIECE using Mo/TiO2 were unsuccessful, a MIECE was successfully made from TiN and TiO2. Test data suggest it behaves as a true MIECE. The morphology and chemistry of the interface between electrode and the alumina were found to significantly influence the charge transfer impedances and ionic conductance (at the surface) events. The surface of beta ″-alumina was mechanically modified using various grit blasting and surface polishing techniques. Once modified, various surface characterization methods were used to understand the morphological and chemical properties of the beta″-alumina surface. These methods included scanning electron microscopy, energy depressive X-ray spectroscopy, X-ray photoelectron spectroscopy, and optical interferometry. Sodium ion conductivity in beta″-alumina measurements were also made. The primary contributor to impedance was surface impurities on the beta″ -alumina. Only after these contaminants were removed was surface roughness found to be influential in charge transfer impedances. These results were confirmed in a series of miniature electrode test cell experiments, and also in a sodium exposure test cell evaluation. more...

The conversion of methane or natural gas into useful intermediate oxidation products instead of complete oxidation to CO$\\sb2$ or partial oxidation to CO is one of the most challenging problems for catalysis research. The relative inertness of the methane molecule towards the activation of its C-H bonds and the tendency for complete dissociation once activation is achieved are the main obstacles to be overcome. This thesis deals with these problems through a fundamental approach and explores catalyst design strategies that would improve the yields of the useful intermediate products. The reaction between methane and gaseous oxygen was investigated in empty Vycor and quartz glass tubes and also over various silica based catalysts. Formaldehyde was the main useful intermediate and smaller quantities of ethane and ethylene were also formed. At low conversions of 0.4% a 93% selectivity towards the useful intermediates was achieved, however, at higher conversion levels the useful intermediates were not preserved (3.9% conversion and 27% selectivity). We suggest that the primary activation of methane is taking place on the silica surface, the oxygen insertion, however, and formaldehyde formation are very likely to include gas phase reactions. Other typical catalyst support materials (alumina, magnesia and titania) led to deep oxidation of methane to CO and CO$\\sb2$. Over the silicas, traces of ethane in the methane feed facilitated the methane activation. When silica was combined with boric acid the further oxidation of the formaldehyde to CO and CO$\\sb2$ was suppressed. Characterization of the used catalyst did not show substantial loss of boron, however, changes in the chemical environment of B$\\sp{3+}$ took place at the drastic conditions of the methane oxidation. Molybdenum based catalysts are commonly used in selective oxidation processes. Through a new method developed in our laboratory, MoO$\\sb3$ of surface area of 40 m$\\sp2$/g and WO$\\sb3$ of 80 m$\\sp2$/g were prepared and characterized. The achieved surface areas were higher than the one of the conventionally prepared materials by as much as an order of magnitude. MoO$\\sb3$ was practically inactive for methane oxidation. Mixed iron molybdates, however, produced promising yields of formaldehyde (83% selectivity at 27% conversion at 973$\\sp\\circ$K). Fe$\\sb2$O$\\sb3$ alone was active, but produced only CO$\\sb2$ at temperatures as low as 673$\\sp\\circ$K. When exposed to methane only, in absence of gas phase oxygen the iron molybdates released some of their lattice oxygen, forming formaldehyde and CO$\\sb2$. Gas phase oxygen greatly improved the rates of formaldehyde, not affecting the rates of CO$\\sb2$ formation. The rates of formaldehyde formation over the non-stoichiometric, molybdenum deficient iron molybdates, were higher than the ones over the stoichiometric iron molybdates by almost an order of magnitude. Analytical electron microscopy of the non-stoichiometric iron molybdates revealed Fe$\\sb2$O$\\sb3$ particles decorating larger Fe$\\sb2$(MoO$\\sb4$)$\\sb3$ particles. more...

Temperature Programmed Desorption (TPD) methods are powerful tools for characterizing gas-solid interactions of importance in surface science, catalysis, and chemical sensors technology. Reaction-Diffusion (RD) systems are dominated by the coupling between kinetic and transport processes. The major objective of this thesis is the development of mean-field models allowing the quantitative treatment of RD systems within the framework of TPD methods. One-dimensional transient models are used to follow the spatiotemporal evolution of concentration gradients within the layers of a single crystal. The strategy implemented eliminates arbitrary decisions regarding the concentration profiles at the beginning of the temperature ramp. Initially we solve the model over the isothermal exposure, and the concentration profile calculated at the end of the exposure is used as the initial condition for the solution over the temperature ramp. The sensitivity of two models to several physical parameters is investigated. One model incorporates bulk and surface states, while the other introduces an additional subsurface state. In both cases Fickian diffusion dominates the behavior of the bulk reservoir, while the surface and subsurface regions are treated as well-mixed domains. The problem involves the simultaneous solution of partial and ordinary differential equations, incorporating rate processes with characteristic time scales spanning several orders of magnitude. We show here for the first time that: (1) Exposure pressure and time do not play a symmetric role within the framework of the definition of a Langmuir exposure for RD systems. (2) The site density of the bulk reservoir and the potential energy diagram (PED) have a non-linear effect on the TPD spectra. Exposure sets at appropriately selected adsorption temperatures in combination with this modeling led to the elucidation of mechanisms underlying the transient behavior of RD systems. Finally, the theoretical framework developed here is applied to the treatment of two specific systems, and modeling results are compared with experimental data. We propose a PED and a set of physical parameters accounting for the TPD spectra of (i) hydrogen on and into Pd(110), and (ii) pre-dissociated hydrogen adsorbed and absorbed into the bulk of Ni(111) single crystal. more...

The priority research areas in heterogeneous catalysis are to study catalyst surfaces while the reaction is taking place and to characterize the actual catalysts at atomic resolution. CO oxidation over Pt/TiOx thin films, as supported model catalysts and sensing elements, are investigated with in situ resistance measurements and ex situ microscopic and spectroscopic characterization. The studies are carried out in a specially designed flow reactor system which can hold thin film model catalysts while simultaneously monitoring the film resistance. The reaction kinetics and the changes of Pt/TiOx thin film resistance during CO oxidation are studied under different O2/CO ratio conditions in a temperature range of 150°C to 375°C. The interaction of CO and O2 with the film surfaces is evaluated by measuring the changes of film resistance and the concentrations of product via on-line gas chromatography. In addition, the effects of the reaction environments on both the composition and microstructure of Pt/TiOx films are characterized with X-ray Photoelectron Spectroscopy (XPS), Rutherford Backscattering Spectrometry (RBS), Transmission Electron Microscopy (TEM), and Atomic Force Microscope (AFM) before and after oxidation. The results show that titanium diffuses through the Pt layer and forms TiOx on the film surface during the oxidation. The surface decoration or metal support interaction behavior on Pt/TiO x thin films for CO oxidation is supported by the kinetic and characterization results. The CO oxidation mechanism on the Pt/TiOx films can not be adequately described by a simple surface reaction model but appears to involve the diffusion of Ti to the film surface. The film resistance serves as an indicator to reveal the dynamic response of the catalyst during the reaction. This study demonstrates that coupling in situ resistance measurements with conventional flow reactor data provides a powerful method for gaining insight into the properties of catalysts under reaction conditions. Also, using thin film model catalysts provides opportunities to fill the gap between single crystal model catalysts and conventional catalysts. The novel approach used in this study is practically applicable and helps to link the macroscopic, practical operation of catalysts with atomic scale structural and mechanistic understanding. more...

The identification of the phase responsible for hydrodesulfurization (HDS) activity has been the subject of extensive research. In this study, model solid state catalysts prepared from elemental starting materials were synthesized, characterized, and then used to desulfurize thiophene at temperatures ranging from 200-400$\sp\circ$C and a pressure of one atmosphere. The results of this work indicate that an increased HDS activity can be correlated with the presence of a poorly crystalline molybdenum sulfide-like phase detected by XRD, HREM, or AEM. The formation of this sulfur-deficient, non-stoichiometric phase could be accomplished by either removing sulfur directly from the catalyst synthesis mixture to yield a non-stoichiometric MoS$\sb{\rm 2-x}$ moiety, or by introducing a transition metal promoter such as Fe, Co, Ni, or Cu into the system. The promoter atoms induced structural changes in the molybdenum sulfide edge planes by effectively scavenging sulfur during catalyst synthesis to form promoter sulfide species, which enhanced the formation of a non-stoichiometric, highly active molybdenum sulfide. This morphological effect was the primary function of the promoter in this system. All model catalysts displayed similar structure in the (0002) basal plane of MoS$\sb2$; however, only the catalytically active samples showed a high concentration of defects and disorder in the (1010), (1011), and (1012) edge planes. The HREM images obtained from these edge planes and their correlation with HDS activity dramatically illustrated the importance of the often-discussed edge plane structure of MoS$\sb2$ and its significance on HDS catalysis. Normalization of the HDS activities for the solid state models and a commercial catalyst with O$\sb2$ or CO chemisorption uptakes suggested that a similarity may exist between the catalytically active sites of these materials. In-situ XPS revealed that increasing promoter atom concentrations resulted in a more complete reduction of the promoter atom; but this shift to lower binding energies could not necessarily be correlated with activity. However, it was observed that the more active catalysts all exhibited the ability to dissociate H$\sb2$ and remove oxygen from their surface. This H$\sb2$ spillover or activation may also influence catalyst performance. more...

Recent advances in microelectronics and silicon processing have been exploited to fabricate miniaturized chemical sensors. Although the capability of chemical sensing technology has grown steadily, it has been outpaced by the increasing demands for more reliable, inexpensive, and selective sensors. The diversity of applications requires the deployment of different sensing materials that have rich interfacial chemistry. However, several promising sensor materials are often incompatible with silicon micromachining and their deposition requires complicated masking steps. The new approach described here is to first micromachine a generic, instrumented, conductometric, microelectronic sensor platform that is fully functional except for the front-end sensing element. This generic platform contains a thin dielectric membrane, an integrated boron-doped silicon heater, and conductance electrodes. The membrane has low thermal mass and excellent thermal isolation. A proprietary selected-area chemical vapor deposition (SACVD) process in a cold-wall reactor at low pressures was then used to achieve maskless, self-lithographic deposition of thin films. The temperature-programmable integrated microheater initiates localized thermal decomposition/reaction of suitable CVD precursors confined to a small heated area (500 $\mu$m in diameter), and this creates the active sensing element. Platinum and titania (TiO$\sb2$) films were deposited from pyrolysis of organometallic precursors, tetrakistrifluorophosphine platinum Pt(PF$\sb3)\sb4$ and titanium tetraisopropoxide Ti(OCH(CH$\sb3)\sb2\rbrack \sb4$, respectively. Deposition of gold metal films from chlorotriethylphosphine gold $\rm (C\sb2H\sb5)\sb3$PAuCl precursor was also attempted but without success. The conductance electrodes permit in situ monitoring of film growth. The as-deposited films were characterized in situ by conductance measurements and optical microscopy and ex situ by electron microscopy and spectroscopy methods. Devices equipped with platinum CVD films showed little chemical response to carbon monoxide and propylene pulses, but exhibited a small response to oxygen and a significant response to hydrogen in vacuum. Finally, a finite difference FLUENT model for low pressure CVD of platinum films was developed and implemented. The process model predicts growth rate, thermal boundary layer profile and an insight into the reaction mechanism. more...

As the environmental protection agency requires increasingly stringent automotive emission controls, new sensor technologies are needed to replace the present oxygen sensors used for air to fuel ratio control. This thesis evaluates the viability of silicon-based microfabricated gas sensors for direct emissions measurements. The chemical sensors consisted of a 1.3 $\rm\mu m\ SiO\sb2/Si\sb3N\sb4$ membrane supported by a 500 $\mu$m silicon rim. A heater and a pair of temperature sensors (TSRs) are suspended from the membrane, while thin films of platinum, gold, and titanium were deposited on top of the membrane for use as gas sensitive films. The structural stability of these devices was tested by exposing them to thermal cycles, pressure shocks, and mechanical vibrations. Several devices that were cycled in temperature over 408,000 times showed remarkable stability up to 650$\sp\circ$C. The sensitivity of various films was measured by exposing the devices to parts per million (ppm) levels of hydrogen, carbon monoxide (CO), or propylene at elevated temperatures in the presence of excess oxygen. Changes in film resistance and temperature were used as measures of sensor response to gases. For example, in response to 500 ppm hydrogen exposure at 300$\sp\circ$C, the Pt/Ti film resistance changed by approximately 10%. The thermal response of some Pt/Ti devices to 2500 ppm propylene was approximately 1.5$\sp\circ$C. Decreased film resistance upon introduction of analyte gases was observed and could be attributed to the increased number of charge carriers or increased charge carrier mobility. Pure platinum and Pt/Ti films were more sensitive to hydrogen than either CO or propylene. The response to all three analytes increased with increasing temperatures. Oxidation of the analyte gases resulted in a measurable temperature increase. This calorimetric effect on Pt/Ti and Au/Ti films was investigated as a method of chemical sensing. The gold containing films showed both thermal and resistive responses to propylene and CO at lower temperatures than the Pt/Ti films. Comparing the relative response of different sensing films and utilizing different sensing methods appears to be a promising approach for gas species differentiation in sensor arrays. more...

The correlation between the microstructure and the activity and product selectivity of catalysts is central to the underst and ing of a working catalyst. This is especially true in supported bimetallic catalyst systems such as Pt-Au and Pt-Sn, two systems which are known to be miscible in the bulk to varying extents. The possibility of the presence of bimetallic alloys is an added complexity to the task of determining the relative distribution of the metallic components in the small metal particles. This microstructural information is vital in the interpretation of the catalytic reactivity trends. A series of bimetallic platinum-gold catalysts supported on high surface area non-porous Aerosil was prepared by the incipient wetness technique. The reducibility of the various metal components in these catalysts was determined by temperature programmed reduction (TPR). Microstructural characterization of the catalysts was performed by using analytical and high resolution electron microscopic techniques. The combination of elemental analysis by EDX and lattice fringe imaging by HREM of individual metal particles was used to determine the presence and location of the various components of this bimetallic system. The kinetic behavior for the n-hexane conversion reaction and the product selectivities towards dehydrocyclization, isomerization and hydrogenolytic cracking were measured. The product selectivity trends in n-hexane conversion can be attributed mainly to geometric effects related to the relative distribution and interdispersion of the platinum and gold atoms in the catalysts. Bimetallic Pt-Sn/Al$\\sb2$O$\\sb3$ catalysts were found to have beneficial characteristics in the activity maintenance for the production of cyclical compounds in the n-hexane conversion reaction. TPR studies indicated that the average oxidation state of tin in these catalysts was Sn(II). HREM and AEM investigations showed that the metal particles in bimetallics with low tin content were elemental platinum surrounded by a patch of ionic tin. However, at higher tin loadings evidence of PtSn alloy formation was found. From these characterization techniques a comprehensive structural model was obtained for these catalysts. An attempt at correlating this structural information with the catalytic trends was made. more...

Bimetallic catalysts have found widespread use in petroleum industry as reforming catalysts. Although the addition of the second metal improves the reactivity of the catalyst, there is controversy as to the exact nature of interactions between the two metallic components which leads to this beneficial effect. Two series of supported bimetallic catalysts with platinum as the primary metal (nominal loading 1 wt %) and varying amounts of second metal (gold or tin) were prepared and characterized. The goal was to understand the nature of interactions between the primary metal and the second metal by studying the reactivity and adsorption behavior of the catalysts. These second metals were used as they have significantly different miscibilities with platinum in the bulk state. The catalytic reactivity was studied using neopentane as the probe hydrocarbon molecule. A detailed catalyst characterization protocol consisting of chemisorption, infrared spectroscopy, and X-ray photoelectron spectroscopy was used. Adsorption studies indicated that addition of gold to Pt/SiO$\sb2$ catalysts led to a progressive decrease in Pt dispersion, and the effect of gold seems to be mainly geometric in nature. The trends in catalytic behavior in the case of the Pt-Au/SiO$\sb2$ catalysts were could be interpreted on the basis of microstructural differences between the catalysts. On the other hand the Pt-Sn/Al$\sb2$O$\sb3$ system turned out to be a much more complex system to analyze. There were significant differences in the oxygen adsorption trends as compared to those observed in the case of hydrogen or carbon-monoxide. The reactivity trends in the case of Pt-Sn/Al$\sb2$O$\sb3$ catalysts indicated that there seemed to be an optimum loading of tin giving desirable selectivity characteristics at reasonable activity levels. The oxidation state of tin in reduced alumina supported Pt-Sn samples as determined by XPS was either Sn (II) or Sn (IV). On the other hand, in the case of silica supported Pt-Sn samples, the tin could be partially reduced to a Sn (0) state. The interactions between platinum and an alumina-stabilized tin species (Sn (II), Sn (IV)) could explain the behavior of Pt-Sn/Al$\sb2$O$\sb3$ catalysts. more...

Microfabricated gas sensors utilizing ultrathin platinum transducing films offer great promise for meeting current challenges to chemical gas sensing. With the research efforts presented in this thesis, considerable advancements have been made toward understanding and optimizing such sensors. The transducing films consist of evaporated, ultrathin (35-100 A) platinum (Pt) films deposited onto metallic titanium (Ti) layers or alumina (Al$\rm\sb2O\sb3)$ layers. Pt/Ti films are studied in microfabricated gas sensing structures and on macrosamples. Titanium may participate in the response of the Pt/Ti films depending upon its oxidation state. Pt/Al$\rm\sb2O\sb3$ thin films are investigated on macrosamples as alternatives to Pt/TiO$\rm\sb{x}$ films used in the microchemical gas sensor. Alumina was chosen as a thermally stable and chemically inactive support. Stability and gas sensitivity of the transducing films are optimized with thermal pretreatments in oxidizing or reducing atmospheres. A range of compositions and microstructures are obtained from one Pt/Ti film. Films pretreated above 600$\sp\circ$C in O$\sb2$ are the most sensitive of the film types investigated. Oxidation of a Pt/Ti thin film at 600$\sp\circ$C results in the preparation of a mixed Pt/TiO$\rm\sb{2-x}$ film with a discontinuous microstructure. Resistance measurements, photoelectron spectroscopies and temperature programmed desorption are used to identify the surface chemical processes responsible for the electrical response of Pt/TiO$\rm\sb{2-x}$ sensing films. Surface reactions involving chemisorbed oxygen play a dominant role in the transduction mechanism of ultrathin platinum films. The films exhibit 10-100% resistance increases upon exposure to O$\sb2$ in the $1\times10\sp{-8}$ to $1\times10\sp{-5}$ Torr range. Films that are pre-exposed to oxygen are sensitive to reducing gases such as hydrogen, propylene and benzene. Resistance decreases in the 20-60% range are observed with reducing gas exposure in the $10\sp{-6}$ to $10\sp{-5}$ Torr range. The mechanistic studies are concluded with temperature programmed reaction spectroscopy (TPRS) experiments involving benzene and propylene on an oxygen saturated platinum foil and 100 A platinum-alumina film. The hydrocarbons react with oxygen at characteristic temperatures to produce CO$\sb2$ and H$\sb2$O. The reaction proceeds with an initial hydrogen abstraction followed by oxidation of the dehydrogenated hydrocarbon intermediates. The dominant parts of the oxidation reactions are the same for the foil and film indicating that alumina does not play a role in the catalytic behavior of platinum-alumina thin films. more...

This research concerns the design, fabrication, performance evaluation, and stabilization protocol of integrated silicon sensors with ultra-thin metal transducing films (100A or less) for gas purity monitoring. The device utilizes a selectively-etched dielectric window which supports a thin deposited metal film with electrodes to sense film impedance. The planar window, 1.3-$\mu$m thick and 1.3 mm on a side, has a buried meandering boron-diffused bulk silicon heater. The heater has a heating efficiency of about 3 mW/$\sp\circ$C in atmospheric air and a time constant of approximately 10 milliseconds. Gases are quantified and identified via ultra-thin film resistance changes that are deconvoluted into separate so-called "thermal" and "chemical" components corresponding to heater cooling and gas adsorption effects, respectively. With temperature programmed desorption (TPD), adsorbed species are identified as they desorb from the surface at characteristic temperatures. In an extension of this technique, a rapid adsorption-desorption equilibrium is established which decreases the detector response time. A custom vacuum test system and computer-controlled data acquisition provides rapid and accurate data collection along with heater and interface circuit control. Thin transducing layers of 35 A of platinum on 65 A of titanium behave as semiconducting, metallic or insulating films depending on pretreatment conditions of gas environment and temperature. The pretreatment stabilizes the films and thereby allows reliable long-term use. The differences in film electrical properties significantly influence gas detection characteristics. One film type detects oxygen to the 10$\sp{-7}$ Torr level, responds to gas in less than one second, and is stable over hundreds of temperature and gas cycles. more...

Thin Pt/TiOx films have proven to be chemically sensitive to their gap environment. To better understand the structure and surface composition of these films, a quantitative determination of grain boundary diffusion of titanium through the platinum layer has been carried out. The accumulation of partially oxidized titanium at the top surface of the platinum layer was measured by x-ray photoelectron spectroscopy (XPS) to determine the diffusion pre-exponential factor and the activation energy for gram boundary diffusion coefficient. For Pt layers thinner than 200 A, there was a thickness dependence on the diffusion kinetics, resulting in activation energies as low as 20 +/- 4 kJ/mol. Above 200 A, the activation energy remained constant at an average value near 118 +/- 15 kJ/mol. The XPS results were validated by conducting in-situ atomic force microscopy experiments on a microelectromechanical structure. The structure is capable of operating from ambient room temperature up to 800°C without damage to the microscope. With this device, topographical images of platinum supported titanium films, have been acquired at temperatures between 25--375°C. In-situ imaging of the titanium underlayer diffusing through the platinum film has been observed at 375°C. Titanium migration to the surface near this temperature is also shown on blanket 35 A Pt/65 A Ti films with x-ray photoelectron spectroscopy (XPS). The conductometric changes in Pt/TiOx films upon exposure to hydrogen and oxygen can be described by a coupled surface reaction diffusion/bulk diffusion model. In this model, the mechanism assumes the disruption of the surface oxygen equilibrium by catalytic reaction with hydrogen. The reaction on the surface provides a driving force for diffusion of oxygen from the layer of TiOx to the surface. The model depends on two parameters: a single parameter to describe the reaction of hydrogen on the surface and a parameter to describe the diffusion of oxygen through the TiOx, layer. The temperature dependence of both model parameters follows an Arrhenius relationship. The activation energies for the reaction and diffusion parameters are 16 kJ/mol and 39 kJ/mol respectively. This work provides valuable information regarding Pt/Ti surface concentrations in thin film chemical sensors, microelectronic applications, and for better understanding the mechanisms of conductometric sensor response. more...